Changing the values of domain names and the need for mapping

JFC Morfin jefsey at jefsey.com
Sat Feb 21 02:01:55 CET 2009


At 20:10 20/02/2009, John C Klensin wrote:
>IDNA is a matter of how _applications_ interpret strings that
>are to be converted into something that can be treated as a
>domain name.   That is why you folks gave it that name.

Applications are out of the end to end scope, but protocols should 
carry all the information applications may need.

To do better than IDN2003 two possibilities:
- more conversion systems, instead of a single one. Not in the 
Charter, yet two special chars (and not accentuated and 
non-accentuated "E" in French which are the upper-cases of 
accentuated "e") mean more conversion.
- drop conversion from the protocol. This is the Charter, but how 
zone managers will publicise the conversion they use?

"DNS" may means: a namespace, a mechanism, operations, a service. 
Name validation in VeriSign DNS service is their legal 
responsibility. AFAIC, they have no RFC, no community consensus, no 
ICANN contract, no WIPO backing to legitimate pure denials of sale. 
Pat, as a smaller zone manager, I would be very interested in your 
legitimacy claim you probably worked with good lawyers - it would help us all.

This raises the question of the IDNA exclusive. Why non-punycode 
performing xn-- domain names could not be used? For example for 
scripts not yet supported by Unicode? or that Unicode supports 
differently from the zone-Manager's own vision?

Why not a privateuse/R&D namespace?

I feel these are issues for a protocol.
jfc



More information about the Idna-update mailing list